WSU Vancouver professor, team research using honey to make advanced computer chips

There’s a chance that a certain sweet ingredient in your cupboard could play a role in the future of computing.

No, seriously: A team of researchers at Washington State University Vancouver has spent the past few years researching how honey can be used to create memristors: tiny devices that process and store data as a key part of neuromorphic computers.

“We are making something great out of ordinary things,” said Feng Zhao, an associate professor in WSU’s School of Engineering and Computer Science. The research was recently published in the Journal of Physics D: Applied Physics, with some funding provided in a grant from the National Science Foundation.

To step back a bit, a neuromorphic computer is a system that operates differently than our everyday phones or computers, which are constructed using what’s called Van Neumann architecture. Van Neumann systems, while they seem fast, said Zhao, are incredibly inefficient in both time and energy.

Neuromorphic computers, however, attempt to mimic how the human brain can control multiple different functions at once, rather than one at a time like the Van Neumann. Additionally, for comparison, the brain uses an estimated 10-20 watts, while regular computers require about 100 watts.

“See how your hand is taking notes based on what I’m saying to you?” Zhao asked. “You’re hearing me, you’re seeing me and your muscles are responding to it.”

In addition to their energy inefficiency, Zhao said, the materials used to create Van Neumann computers are neither sustainable nor renewably produced.

“All computer and chips are developed with silica. They’re incredibly wasteful in production and don’t biodegrade in any real, tangible amount of time. Nothing can be done with them when they’re used up,” Zhao said. “Honey dissolves in water, it’s renewable, biodegradable. It could be the solution to two major global issues of both energy and waste.”

In the past, Zhao has worked with other, similar organic compounds like sugars and aloe vera. Honey, he said, is the most promising because of a low moisture concentration that makes it difficult to harbor or collect bacteria. Not only that, but once they’re used up, they can be easily disposed.

Producing each piece

Each individual memristor is about 100 micrometers in size, about twice the width of a human hair. Made up of a team of graduate students, Zhao’s lab presses the honey into thin slivers – about 10-20 micrometers in width – and places them between two slivers of metal that serve as conductors.

The memristors are then hooked up to a probe station that tests their ability to mimic the work of synapses.

Since the devices are so delicate and small, the process to create a single memristor takes approximately two weeks. To build an entire chip, they’d need millions of these memristors. Not only that, but they’d need to be significantly smaller in size.

Major companies like IBM and Intel are working to create neuromorphic computers, Zhao said, but are still using unsustainable materials like silica to build them.

“Compared to silica, our processing – if production is possible – is much simpler and comes at a lower cost,” Zhao said.

The honey isn’t any kind of special honey, either.

“You can get it at Walmart. Any grocery store, really,” said Brandon Sueoka, a graduate student studying electrical engineering who’s worked in Zhao’s lab since 2020.

Since it wouldn’t be feasible to produce millions of these memristors on their own, Zhao said the next steps would be to continue to optimize the chips by testing different factors in production such as the temperature that the devices are heated at, how long, and what metals work best as electrodes. On Wednesday, the device that Zhao had on display used tiny pieces of silver as the top electrode; previously, they had used aluminum.

Once optimized, they would look to find a partner for mass production.

More information on Zhao’s lab and other research in applied engineering at WSU Vancouver is available at labs.wsu.edu/feng-zhao/research/.